When Artificial Intelligence Interprets New Ideas Through Traditional Baseball Thinking

Major League Baseball has never had access to more information.

Organizations invest heavily in technology, performance science, biomechanics, motion capture, tracking systems, data analytics, and increasingly sophisticated player development infrastructure.

Now, Artificial Intelligence is opening another door.

The possibilities appear enormous. AI systems can process vast amounts of information, analyze video, recognize patterns, reproduce coaching knowledge, and generate recommendations. They can potentially make information more accessible and help organizations explore new approaches to player development.

I am hearing that some professional baseball organizations are actively exploring AI in search of a competitive advantage in Player Development.

In some conversations, the ambition sounds almost like a search for the Holy Grail.

And that raises an important question:

What happens when we ask AI to find new answers while its interpretation of those answers remains heavily influenced by established ways of thinking?

Because there is a fundamental difference between finding information, reproducing knowledge, and actually understanding what that information represents.

And in player development, that difference matters.

THE PROMISE OF AI — AND THE QUESTION BEHIND IT

Imagine an AI system trained on an enormous collection of baseball knowledge.

Books, research papers, coaching philosophies, instructional videos, biomechanics, performance data, and years of professional experience.

A coach uploads a video of a player. The system analyzes the movement, identifies patterns, compares the results with its knowledge base, and produces recommendations.

For pitchers, it might evaluate throwing mechanics, delivery patterns and movement characteristics.

For hitters, it might analyze the swing, sequencing, rotation and bat path.

The same principle could extend to infielders, outfielders, baserunners and virtually every other aspect of the game.

On the surface, this sounds extremely promising.

But there is a question we should not overlook:

Does the system actually understand the individual, or is it interpreting the individual through the knowledge and assumptions it has been given?

Those are not necessarily the same thing.

An AI system can reproduce a coaching philosophy with remarkable consistency. It can recognize movement patterns, compare them with established models, and generate technically convincing explanations.

But reproducing an existing understanding does not automatically mean the system can recognize when that understanding is incomplete.

And it certainly does not mean that the system has identified the individual behind the movement.

WHEN AI FINDS INFORMATION BUT MISSES THE POINT

I recently experienced something that illustrates this problem.

Someone who had heard about my work decided to use AI to learn more about me and MotorBall.

The system found information. It was able to retrieve fragments of my work, including material published on my Substack.

However, the resulting interpretation was incomplete and, in important respects, fundamentally incorrect.

MotorBall was compared with approaches that are fundamentally different from what it actually represents.

Why?

Because the information was being interpreted through familiar concepts and established baseball frameworks.

Instead of understanding the underlying principles, the system appeared to translate what it found into concepts it already recognized.

That distinction is important.

MotorBall translates the principles of the official ActionTypes framework into baseball. It looks at the individual’s natural motor organization and how that organization relates to movement, performance, communication and development.

It starts with a different question.

Not simply: What does the player do?

But: How is this individual naturally organized to move?

When an unfamiliar concept is interpreted through an established framework, its original meaning can change.

The result may sound convincing. It may contain familiar terminology and technically plausible explanations.

But a convincing explanation is not necessarily an accurate understanding.

And this is not simply a question of whether information about a particular approach is available online.

Information can be available without the framework behind it being properly understood.

Finding information is not the same as understanding the framework behind it.

THE MLB PARADOX

This brings me back to the growing interest in AI within professional baseball.

MLB organizations operate at an extraordinarily high level of sophistication. They have access to some of the best technology, research, personnel, and performance infrastructure in professional sports.

Yet the introduction of AI creates a new challenge.

How carefully are organizations evaluating the knowledge, assumptions, and underlying frameworks that these systems use to produce their answers?

AI can process enormous amounts of information. But the quality of its output depends on more than the quantity of information available.

It also depends on the quality of the source material, the context in which that material is interpreted, the assumptions built into the system, and the way its conclusions are validated.

If the information is incomplete, the interpretation can be incomplete.

If the underlying assumptions are wrong, the system can reproduce those assumptions at scale.

And if a new concept is interpreted through an established framework that does not accurately represent it, the resulting explanation may take us further away from the original idea rather than closer to it.

This is not an argument against AI.

It is an argument for being more critical about what we expect AI to understand.

The question should not only be what a system can produce.

We should also ask what it actually knows, what it assumes, and what it might be missing.

THE PLAYER BEHIND THE MOVEMENT

This becomes particularly important when we talk about individual player development.

Every player has a movement history.

A pitcher’s delivery, a hitter’s swing, an infielder’s footwork, an outfielder’s movement toward a fly ball, or a baserunner’s acceleration are all visible expressions of movement.

But what we see today is not necessarily a direct representation of how that individual is naturally organized to move.

Years of coaching, technical adjustments, training methods, physical adaptations, and individual experiences can all influence the movement we observe.

A player may have developed a particular movement pattern because it was taught, because it produced a desired result, because it helped compensate for something or simply because it became familiar over time.

That movement can be measurable.

It can be repeatable.

It can even be highly effective.

But identifying the movement is not the same as identifying the natural organization behind it.

This is a distinction that deserves much more attention in player development.

We have become increasingly sophisticated at measuring what players do.

We can quantify movement, evaluate outcomes, identify patterns, and compare performance.

But how do we determine whether the movement we see reflects the player’s natural organization or something that has been acquired over years of instruction and adaptation?

And more importantly, how do we know whether our interpretation is correct?

THE DIFFERENCE BETWEEN ANALYZING MOVEMENT AND UNDERSTANDING THE INDIVIDUAL

Consider the possibilities.

An AI system analyzes a pitcher’s delivery and identifies a specific mechanical pattern.

It analyzes a hitter’s swing and identifies a particular sequence of movements.

It evaluates an infielder’s fielding mechanics or an outfielder’s movement efficiency.

It can potentially compare these patterns with thousands of examples and generate recommendations based on established knowledge.

That can be valuable.

But what if two players demonstrate similar movement patterns while being naturally organized in fundamentally different ways?

What if two players achieve similar outcomes through different movement organizations?

And what if a movement that appears mechanically unusual is actually consistent with the individual’s natural organization?

The visible result alone does not necessarily answer those questions.

This is where I believe the next important conversation in player development needs to take place.

MotorBall approaches the individual from the perspective of natural motor organization. Through the principles of ActionTypes, it seeks to identify an individual’s Motor Signature — their Natural Blueprint — and understand how that organization relates to movement and performance.

The purpose is not to replace biomechanics, tracking technology, video analysis, or other performance tools.

It is to introduce another layer of understanding.

Because measuring the movement is not the same as identifying the individual.

And more information does not automatically produce a deeper understanding of the person we are trying to develop.

FIVE QUESTIONS THE INDUSTRY SHOULD BE ASKING

Before organizations place too much confidence in AI-generated player development recommendations, I believe there are some fundamental questions worth asking.

1. What are you actually measuring?

Are you identifying observable movement characteristics, mechanical patterns and performance outcomes, or are you claiming to identify something beyond what can be directly observed?

2. How do you distinguish natural organization from acquired movement?

How do you determine whether a player’s current movement reflects their natural organization or years of coaching, technical adjustments, and adaptation?

3. How do you validate your interpretation?

What establishes that the system’s explanation of a player’s movement is actually correct, rather than simply consistent with the information it has been trained on?

4. How do you rule out misidentification?

If a system assigns a player to a particular movement category or recommends a specific approach, how do you establish that the underlying identification is accurate?

5. What does individualization actually mean?

Are we genuinely adapting development to the individual, or are we simply using more sophisticated technology to deliver variations of recommendations based on established models?

These are not questions designed to dismiss technology.

They are questions designed to establish what the technology is actually capable of telling us.

THE RISK OF AUTOMATING AN INCOMPLETE UNDERSTANDING

There is an enormous difference between using AI to make existing knowledge more accessible and using AI to discover something we do not yet fully understand.

The first is already a powerful application.

The second is a much more demanding challenge.

If we feed a system established coaching philosophies, research, instructional material, and performance data, we should expect it to become increasingly capable of reproducing and applying that information.

But if the underlying knowledge does not adequately account for individual differences, the system may simply reproduce those limitations.

And when that happens, the technology does not solve the problem.

It can make the problem more efficient, more consistent, and potentially much more widespread.

Automating an incomplete understanding does not make it complete.

It simply allows that incomplete understanding to be reproduced at scale.

This is particularly relevant when the objective is to develop individual players rather than simply classify, compare or optimize observable performance outcomes.

The challenge is not just to tell a player what they are doing.

It is to understand who that player is, how they are naturally organized to move, and what that means for their individual development.

That applies to pitchers and hitters, but equally to infielders, outfielders, baserunners, and every other player on the field.

The position changes.

The task changes.

The movement changes.

But the individual behind the movement remains the starting point.

WHAT IF WE ARE ASKING AI THE WRONG QUESTION?

I am not questioning the potential of Artificial Intelligence in baseball.

Quite the opposite.

I believe AI could become an extremely valuable tool in player development.

But I also believe the industry needs to be careful not to confuse the ability to process information with the ability to understand the individual.

AI can help us access more knowledge.

It can help us analyze more information.

It can help us identify patterns that might otherwise go unnoticed.

But the real question is whether it can distinguish between what a player does, why that movement has developed, and how that individual is naturally organized to move.

Those are different levels of understanding.

And we should not assume that progress at one level automatically means progress at the next.

Perhaps the next competitive advantage in player development will not come from finding an AI system that knows more baseball than everyone else.

Perhaps it will come from asking better questions about the player in front of us.

Because the real challenge isn’t teaching AI to repeat everything we already know.

It’s recognizing what we still don’t understand about the individual we’re trying to develop.